• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Ryzen 5 3600X OCing

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Merlins Beard

Registered
Joined
May 15, 2020
Hello guys,

Finally upgraded to AM4, now...... from what i can tell here a few things are different on Ryzen i have looked at a few guides and all of them i have seen talk about undervolting Ryzen for performance gains rather than adding voltages, leaving me hard pressed on how to approach this any input on info to help me get my head around this if A2 B2 ram configs wasnt weird enough lol.
 
At stock (auto) voltages it goes anywhere up to 1.5v but it's only for single core short bursts, for all core manual OC you should aim for 1.3v-1.325v max to avoid degradation. There's a lot of confusion about this still, but the Stilt stated that he nearly killed a 3700x at 1.33v for putting him at high loads (FM3, 90c+) for long periods of time.
 
Lol rip that dude sounds like he needs better cooling, you have me a nice solid idea of a starting point though ! My 3600X refuses 4.4GHz unless it’s like 1.45v plus but 4.3GHz it’s okay at like 1.3V minus (still testing and lowering) Ryzen is bizarre
 
Nah even on a custom loop my 3700x will hit 90c with PBO enabled. Reddit is telling people max of 1.2v but that seems pretty low. Honestly I would suggest just running the thing completely stock unless you do a lot of multi-threaded work. For gaming the stock settings are almost always best. On my card enabling PBO, 10x scalar, max limits increased temps by 10c but did not improve max single core clock (actually it was about 25MHz lower) and actually lowered the Cinebench R20 score (probably because heat, not power limits, is the primary limiting factor of the boost algorithm, especially for the 105W TDP CPUs).

Edit: my CPU was able to run CB R20 at 1.275v and 4.3GHz. I didn't long term test it though.
 
After two days of Messing around with the OCs i kinda agree with you Zerileous i have only managed to increase my voltages and Temps i cant OC past 4.3Ghz which is in the boost already lol ! rip i expected alot more from the R5 3600X TBH
 
Why, it beats the 8700k at productivity and competes with it in gaming at lower speeds, lower TDP and half the price (last time I checked). The 3600 non-X is being called the best overall chip on the market pretty much since it came out.
 
hmmm fair doos, im just soo use to AM3+ being able to make alot more Clock speed gains over the boost lol but i love how the R5 3600X performs.... i just have to get use to my MSI Bios lots of new things added in here but i am in love with the M.2 :)
 
hmmm fair doos, im just soo use to AM3+ being able to make alot more Clock speed gains over the boost lol but i love how the R5 3600X performs.... i just have to get use to my MSI Bios lots of new things added in here but i am in love with the M.2 :)

This will be the new normal for all CPUs and gpus as they all go onto smaller manufacturing processes. This is slightly speculation of course but as all the manufacturers shrink the process down they will have to reduce clock speeds. You couple that with the boost technology that all of the manufacturers are now using which gets so much performance out of the chips without much intervention from us and basically the overclocking is done for us. Also in some situations it is best to leave the boost technology in place rather than an all core Overclock, as the higher boost to a single core is advantages in some tasks. Heavily single threaded games for example prefer it to boost that one core rather than a lower all core overcloc.


 
In a way I guess the shrinking fab size is limiting per core frequency as the die is smaller and can't dissipate heat to the IHS as well. But that is somewhat offset by allowing the need f or less voltage since the transistors are closer together. I think the big issue is the proliferation of cores and cache and the thermal challenges that produces.
 
After two days of Messing around with the OCs i kinda agree with you Zerileous i have only managed to increase my voltages and Temps i cant OC past 4.3Ghz which is in the boost already lol ! rip i expected alot more from the R5 3600X TBH
all core boost is 4 ghz(?) in that cpu...but 2 cores are 4.4. I
 
After some more testing i have found i was able to reduce the voltage a fraction over what the CPU was drawing on auto with boost, stable at 4.3Ghz @ 1.415 Volts over the 1.455 v on auto lol
 
I would be cautious running a fixed overclock at those voltages. I've seen numbers as low as 1.2v thrown around as "max safe" although I think 1.3v is a more reasonable number. The stock boosting algorithm takes into account a whole lot more than voltage and temperature, and it will boost using voltages on auto that will cause degradation in a short amount of time when set as a manual OC.
 
well after a few tests on stock voltages rarely climb down from 1.4 according to AMD they consider 1.4 to be a reasonable voltage my motherboard and alot of info i have found consider anything above 1.45v excessive but as i said earlier on stock the cpu draws more than 1.43 volts consistantly
 
Okay. Some people have run at 1.4v since release without issues. Others have had their CPUs degraded and unable to run stock after just weeks at that voltage. It's your hardware at the end of the day. However it is not appropriate to state that just because the stock behavior uses those voltages you can for a fixed overclock. The stock boosting algorithm factors in temperature and current in a more granular way than you can. To demonstrate, run something more demanding like Cinebench R20 or prime95 small FFTs and you will see the voltage likely drop below 1.3V with stock settings.
 
As I said on my 1st post - “At stock (auto) voltages go anywhere up to 1.5v but only for single core short bursts (its within spec), for all core manual OC you should aim for 1.3v-1.325v max to avoid degradation.”

This is after testing done by The Stilt, Der8auer and other known overclockers and reviewers using FIT, it’s an quick and easy google search. Also, as far as I know, AMD never stated that 1.4v is OK for all core OC on Ryzen 2.

“Enable PBO with stock PPT, EDC, TDC, leave the other CPU settings at auto. Then run Prime95 small FFT and watch the SVI2 TFN sensor. Whatever that reads, that is your fitness voltage. The average FIT voltage is 1.32v, but I have seen reports of chips as low as 1.25v. The average person shouldn't risk their chip for an extra 50-100MHz. If you want safe FPS gains, optimize your memory + infinity fabric.”
 
My apologies, i just found it hard to believe that the R5 3600X is uncapeable of handling 1.4v+ ~ 1.45V for sustained periods i did read from AMD that ideal voltage is 1.3v even 1.4v is acceptable on an OC. I do see what you are saying Kenrou on auto watching the voltages on HWmonitor and aswell on Ryzen Master the voltages to fluctuate from core to core but most of the time i see it spend more time in the 1.4v+ range than its short bursts at 1.3v and below range but with that said after looking at the results i was getting at what i would personally consider a safe voltage range, the performance is a fraction less than the performance i was getting at stock and i noticed there is more of a variation with core to core performance which i find quite interesting. So as i said i kinda agree with you zerileous i might aswell leave it stock lol.

as for the infinity fabric not alot in my research and reviews mentioned anything about it when i was buying my upgrade as far as i can tell its tied to the frequency of the RAM if i am not mistaken, can anyone give me some material ? i am rocking Corsair RGB vengance pro @ 3200Mhz is there much i can do to tweak things like timmings, speed, and "infinity fabric"
 
The issue with infinity fabric is in relation to RAM frequency. When you get above about 3700 mhz on the RAM with Ryzen 3000 series CPUs the RAM frequency decouples from the infinity fabric such that the ratio of the two frequencies is no loner 1:1. That produces a performance hit that isn't offset unless you can get the RAM up to 4000+ mhz. You are not anywhere near that problem.
 
No apologies needed, we're just trying to help you not to damage your hardware. Older versions of Ryzen indeed could handle more voltage for an all core OC, but the die change to 7nm also changes what it can handle. Ryzen 1*** were on 14nm, Ryzen 2*** were on 12nm (correct me if i'm wrong) and Ryzen 3*** are on 7nm. Silicon Lottery also has a lot to do with it, you might get a chip that can handle 1.35v and one that can only go as far as 1.25v, my 3600 FIT was 1.318v which did 4.2ghz all core.

Day to day (24/7) we rarely push the CPU even when gaming, for example, the only time my 3800x went above 70% usage outside of a stress test was loading Mass Effect Andromeda for the 1st time, and it likely won't do it again until I restart editing videos. The auto settings push 2 cores to high voltage when needed but as I said these are for very short duration in which the voltages don't matter (up to ~1.5v is within spec) unless you have crappy cooling. Ryzen Master will tell you which cores are best.

Buildzoid (motherboard reviewer) killed he's 3700x running 1.375v.

Here's an example of degradation on a 2700x -
 
Thank you very much for your input guys this is very valuble to me ! i was so use to the AM3+ platform not realising when i upgraded just how different it really is on the FX 8350 for example you use to be able to just whack on all core OCs and give it voltage under 1.5v, but yeah in terms of this im going to leave my 3600X at stock i think but delve in to this infinity fabric and focus more on the ram because it sound like the ryzen architecture really benefits from RAM performance most.

anything anyone can give me in terms of guidance or guides ect to improve the IF and ram i will admit i am some what of a greenhorn with OCing ram i had a go with memory try it on my MSI B450 Motar max but nothing seems to work (as expected with auto OCs never work lol)

Also to note is Ryzen Master worth using as i usually prefer bios OC over a program based system it just sounds like a rebranded AMD Overdrive
 
Last edited:
Ryzen Master: if you don't want to reboot to make changes, it works fine. But most prefer a BIOS OC. Granted I have heard of people overclocking individual cores using Ryzen master, then a program called process lasso to control process affinity to make a game or whatnot use the highest clock core(s).

1.325v was the original number that was passed around for safe voltage in an all core overclock, then some people reported degradation after 6 mo. or so at that voltage. The fit voltage thing game about. It's something I need to read more about because of course thermals have something to do with it. Also for example, if I enable PBO (note if you just max the limits, it doesn't work RN, for me using the "105w" TDP numbers (PPT 142, TDC 95, EDC 140) is plenty to keep my CPU "65w" from hitting the limits adjusted by PBO). Note the quotes because at this point TDP has very little to do with how many watts the system utilizes. If I do this my CPU will exceed 90c tdie and voltage will roll back to 1.2v or less running Prime95 Small FFT. Conversely with a fixed OC of 4.3GHz at 1.275v set (closer to 1.24v SVI) temps stay under 85c running Small FFT. Doesn't make any sense to me.

edit sorry I got off on a tangent. With 3200MHz memory, simply enabling XMP will get you 95% of the performance as additional tweaking. Up to 3600MHz memory, all you have to do is adjust the memory speed and the infinity fabric will automatically follow. After that it depends on your motherboard, but you will either have a fabric clock expressed as MHz (1600MHz, 1800MHz, etc): FCLK, or it will be expressed as a ratio of either 2:1 or 1:1. For a straight forward explanation:
, to get into the weeds / you need to take a nap:
 
Back